The hidden costs of cloud computing are similar to an iceberg; the real threat lurks below

New features and services emerge every week on the cloud. As platforms continue to grow, it becomes increasingly difficult to accurately measure costs.

At the same time, people think cloud costs can be broken down into compute, data, and bandwidth. They’re not wrong, these are the key components of your bill. But, if you dig a little deeper and take a more granular look, you’ll discover there’s much more to it than that.

What are the hidden costs of cloud computing?

They say a picture's worth a thousand words. If you want to get an idea of just how complex cloud costing is, look no further than this diagram from AWS.

A diagram illustrating the complexity and hidden costs of cloud computing

1. Compute

Cloud platforms use autoscaling to dynamically adjust resources in line with current demand. There’s no fixed rate. You pay less during periods of low traffic and more during periods of high traffic. This has given rise to a common misconception that compute is simply the amount you pay each month for every virtual machine in your estate. Unfortunately, it’s not that simple.

For example, it’s not only the server costs that increase during periods of high demand but the logging charges as well. You also have to factor in the trade-off between security and cost. In my experience, it’s impossible to have a system that’s both 100% secure and inexpensive to run. That’s not to say that an optimised cloud environment is vulnerable to attack. The reality is that the more secure your cloud, the more expensive it’ll be.

2. Data

Database engines offer a broad range of services—from high availability and multi-region to always-on. Each of these systems come with additional costs, but this is often overlooked, and businesses continue to pay for the same services without considering their requirements.

This comes back to a lack of visibility. Many businesses don’t know what data they have or where it’s stored. I’ve worked with clients who keep gigabytes of obsolete data in expensive data warehouses, like Amazon’s Elastic File System. It’s only an extra $10 a month, so why go through the hassle of finding an alternative solution? But over one, three, five years, it quickly adds up.

Critically evaluate your database. Decide what you should keep and what you can get rid of. If you need to preserve historical records for regulatory purposes, consider moving them to cold storage. It’s more cost-effective and helps you better organise your data.

Discover how to adopt and test new tech without increased risk. Download the  whitepaper.

3. Bandwidth

Accurately calculating bandwidth costs is notoriously difficult. Even Amazon hasn’t worked out a solution yet and honestly, I don’t think it can be done. There are just too many variables.

For example, many businesses don’t have a clear idea of their monthly bandwidth requirements or how periods of high usage differ from normal activity levels. When they get their bill at the end of the month, they’re often shocked to see just how much they’ve spent on outgoing data charges. Data transfer rates also differ from region to region. Brazil, for instance, charges 30-40% more than the UK and the US. So, if your servers are connected with a third-party service located in a different availability zone, your cloud costs will change.

With so much to consider, many Managed Service Providers omit bandwidth from their cost optimisation services. At the very least, they make it clear to the client that they’re only able to provide a broad estimate.

How to spot the hidden costs

  • Set a clear budget and monitor it carefully as the month progresses

Let’s say you spend £1,000 a month on cloud services, but you’ve reached 50% of your budget in the first week. By scrutinising your bills, you can quickly see when something’s not right and investigate. You might need to rethink your budget or, perhaps, simply re-spec one of your instances. The point is to keep track of your costs.

  • Interrogate and critically review your cloud bills

I recommend doing this once a month if you iterate regularly. Look at your resources and compare them with the previous month to see what, if anything, has changed and look for new trends. For example, you might find that there’s a spike in demand on Saturday mornings that you need to factor into your budget. If you use AWS, one of the best things you can do is tag your resources. This will help you break down your bill beyond the individual services you use and gain a more granular view of your costs. Just remember that tagging only works if you do it consistently.

  • Explore your pricing options

Businesses waste huge amounts of money every year on underutilised and obsolete services. A savings plan and or reserved instances that give you a price point reduction if you commit to a specific vendor might be a more cost-effective option than your current solution—as long as you aren’t put off by vendor lock-in.

Why are cloud costs so poorly understood?

Too many people accept bills at face value. Whether it’s because they come from a traditional data centre background and are used to seeing a handful of list items on their bill or assume that, if they’re paying something, they must need it, people underestimate the complexity of the cloud.

To spot those hidden costs and optimise your environment, the best bit of advice I can give you is to develop a clear understanding of your cloud environment. The more insight you have, the easier it is to keep on top of your costs.

Jumpstart tech investment and ai pilot methodology